Alzheimer's disease (AD) is the most common age-related dementia, and is rapidly becoming a major public health problem. At present, there are no cures or disease-modifying therapies for AD. Without options to cure or even delay the onset of AD, the economic costs, healthcare burden, and impact on afflicted individuals, their families, and society will be devastating. The overall goal of this project is to examine the effect of aerobic exercise on central metabolic pathways using metabolomic and proteomic markers. We will leverage blood samples from 350 older adults with MCI enrolled in the funded ADCS-funded EXERT study. In this study, the two primary outcome measures will be quantitative levels of 24 metabolites and 21 proteins from previously published panels. The main independent measure will be exercise condition (aerobic exercise, stretching, sedentary) in the EXERT study. For the primary analyses, this will be a repeated measures, within-and- between subjects design examining exercise group effects on the metabolomic and proteomic measures. In a model development arm, we will integrate the metabolomic and proteomic measures in a putatively richer and more specific multi-omic model which may be more sensitive to exercise effects than the already established individual panels. Using the multi-omic and individual panels, we will also attempt to predict exercise effects for individual participants from baseline data. We will examine change in the metabolomic, proteomic, and multi- omic measures for other EXERT outcomes including CSF measures of amyloid and tau, brain atrophy from imaging, and other blood measures. Finally, in an exploratory analysis, we will perform untargeted metabolomic quantification of ~4700 metabolites to determine if more specific metabolomic panels may be identified. Our team is comprised of experienced basic and clinical scientists, clinical trialists, and biostatisticians with the resources to carry out this project. The results of this trial will elucidate metabolic mechanisms of exercise as a treatment and possibly prevention of Alzheimer's disease. The multi-omic panel and metabolomic panel from the untargeted metabolomics may provide new avenues for detecting Alzheimer's disease and for monitoring treatment effects in future clinical trials.